Campi Flegrei

25 downloads 0 Views 3MB Size Report
Feb 24, 2017 - ted gases from the fumaroles was carried by Aiuppa et al. ...... A.L. Deino, G. Orsi, S. de Vita, M. Piochi, The age of the Neapolitan Yellow Tuff.
International Journal of Mass Spectrometry 415 (2017) 44–54

Contents lists available at ScienceDirect

International Journal of Mass Spectrometry journal homepage: www.elsevier.com/locate/ijms

Real-time quadrupole mass spectrometry of hydrothermal gases from the unstable Pisciarelli fumaroles (Campi Flegrei): Trends, challenges and processes Alessandro Fedele a,∗ , Maria Pedone a , Roberto Moretti b , Thomas Wiersberg c , Renato Somma a , Claudia Troise a , Giuseppe De Natale a a

Istituto Nazionale di Geofisica e Vulcanologia, Sez. di Napoli – Osservatorio Vesuviano, Italy Università degli Studi della Campania “Luigi Vanvitelli”, Dipartimento di Ingegneria Civile, Design. Edilizia e Ambiente, Aversa, CE, Italy c Helmholtz-Zentrum Potsdam – Deutsches GeoForschungsZentrum GFZ, Germany b

a r t i c l e

i n f o

Article history: Received 9 September 2016 Received in revised form 7 February 2017 Accepted 16 February 2017 Available online 24 February 2017 Keywords: Quadrupole mass spectrometer Fumarolic gases Hydrothermal systems Continuous geochemical monitoring Campi Flegrei

a b s t r a c t Volcanic gas sampling and post-collection chemical determination in a laboratory may preclude any real-time continuous monitoring of volcanic activity. We describe the development, and show the advantages, of a system used for the continuous monitoring of fumarolic gases discharged from the Pisciarelli site (Campi Flegrei, Southern Italy) based on a commercial quadrupole mass spectrometer (QMS-301 OmnistarTM ). Although numerous technical problems were addressed due to the ephemeral nature of the emission point and the harsh environment, we also report measurements of the chemical composition of the major gas species emitted from the fumarole for two different periods (in 2009 and 2012). The CO2 /H2 S, H2 S/H2 , He/CO2 and CH4 /CO2 molar ratios were investigated in order to detect magmatic and/or hydrothermal components in the system, while the N2 /O2 ratio was adopted to infer other non-volcanic processes, such as air contamination and mixing with polluted surface waters. The presented methodology allows continuous gas sampling and provides the first evidence of short-term gas variations not available by direct sampling, which is often impractical and hazardous. Compared to the current techniques that are used worldwide for the full characterization of gaseous emissions, i.e. chemical analysis of traditional soda-filled under-vacuum bottles and MultiGAS surveys (laboratory-based and in situ, respectively), QMS-based monitoring is complementary and, in prospect, an alternative. With our method, the geochemical monitoring benefits of the real-time analysis for high sampling rates that can be made comparable to the continuous measurements of geophysical networks. This allows a better understanding of hydrothermal features, particularly of chemical fluctuations occurring on the very short-term, which is fundamental for the evaluation of the evolution of unrest episodes at Campi Flegrei, one of the most hazardous volcanic areas in the world. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Several efforts have been made in geophysical and geochemical monitoring to interpret, and possibly predict, the evolution of volcanic activity. Typically, geophysical methods target the depth and size of the magmatic/hydrothermal source by inverting seismic features and ground deformation data from widespread acquisition ground networks, such as the seismic and GPS ones, in some cases integrated by satellite interferometry for ground displacements over wide areas; geochemistry can be either based on in situ site-specific measurements (e.g. fumarolic gases sampling and sub-

∗ Corresponding author. E-mail address: [email protected] (A. Fedele). http://dx.doi.org/10.1016/j.ijms.2017.02.006 1387-3806/© 2017 Elsevier B.V. All rights reserved.

sequent laboratory analysis, thermal profiling in boreholes), or widespread data acquisition covering wide surfaces (e.g. soil diffuse degassing). Isolated point measurements, and subsequent laboratory analyses, yield the full concentration of fluids discharged at specific peculiar sites (fumaroles) and allow interpretation of the pressure and temperature conditions of actively degassing hydrothermal reservoirs by identifying the conditions yielding the last gas reequilibration prior to the outlet discharge. Such an approach to the measurement of intensive parameters of the hydrothermal system must be necessarily complemented by the measurements of extensive quantities, such as the extension of the anomalous (i.e. exceeding the background value) degassing areas, and the fluxes of mass discharged at surface.

A. Fedele et al. / International Journal of Mass Spectrometry 415 (2017) 44–54

Volcanic eruptions are frequently preceded by variations in the chemical and isotopic composition of fluids discharged from fumaroles, hot springs and also from diffuse soil emissions making observations of waters and gases basic tools to understand the magmatic-hydrothermal systems beneath active volcanoes [1–4]. In particular, diffuse CO2 emissions have been quantified by soilgas surveys of hydrothermal systems at volcanoes in a quiescent condition and, recently, more information have been obtained on fumarolic CO2 emissions [5]. However, high-precision geochemical discrete sampling and measurements suffer limitations, essentially due to the necessity of accessing the point source of gas emission, avoiding any secondary phenomena (such as component removal, particularly water, or contamination). Collection of fumarolic gases is – depending on logistics and the state of volcanic activity – mostly performed discontinuously with time intervals between consecutive sample collections ranging from days to weeks or even months between consecutive sample collections [6]. Sampling frequencies are too low to allow an efficient comparison between gas data and geodetic and seismic information from continuous geophysical monitoring. Furthermore, direct sampling poses dangers when getting close to active fumarolic sites. Steam, from both magmatic and superficial heated up sources (such as overlying lakes or groundwaters), is the most common volcanic gas, but other gases, often very toxic, are emitted during eruptive events and may impact on human health. CO2 and H2 S, particularly, are heavier than air and may pool at ground level and result in asphyxia [7]. Another disadvantage of the direct sampling is that it requires instrumental analysis in the laboratory, which is often performed many days later making it unsuitable for tracking real-time variations on site. Because of these difficulties, the volcanic CO2 flux inventory (claimed as one of the most reliable gas precursors to an eruption [8]) remains sparse and incomplete for most of the active volcanoes on Earth [9]. Moreover, its estimate is further complicated by the high background atmospheric concentration, which, inexorably, can affect all measurements carried out by remote-sensing techniques [8]. High-frequency measurements of volcanic gases are a priority for obtaining a better understanding of the precursors of eruptive activity and unrest, but because increasing the frequency of on-site discrete sampling in such environments entails high risks and costs, in the last decades, automatic systems have been developed (MultiGAS) [10–13] and several novel techniques are still in experimental phase (Tunable Diode Laser TDL [5]; Differential Absorption Lidar DIAL [8]) to become a strong tool in geochemical monitoring. Few applications of continuous fumarolic gas monitoring systems based on quadrupole mass spectrometer have been so far discussed in literature. Continuous measurement of dissolved gas concentrations using a quadrupole mass spectrometer (QMS) was described by Takahata et al. [14]. Faber et al. [15] reported on a system which analyzed fumarolic gases pumped through a pipe to a remote station where several instruments (e.g. gas chromatograph and mass spectrometer) monitored various physical and chemical parameters. In situ mass spectrometer measurements of volcanic gaseous emission were conducted at several fumarolic sites using a commercial quadrupole mass spectrometer by Diaz et al. [16] at Poas (Nicaragua) and Hawaii Volcanoes. It is worth noting that QMS represented an important advancement compared to other techniques for continuous gas monitoring which have been applied on several volcanoes. Earlier attempts by Le Guern [17] were based on a portable gas-chromatograph to determine in real-time the full gas speciation at Vulcano Island (Italy), Kilauea (Hawaii, U.S.A.), and Merapi (Java, Indonesia) volcanoes. Similarly, Zimmer and Erzinger [18] and Zimmer et al. [19] showed data from a gas chromatographbased system operating continuously at the summit of Merapi Volcano. By means of specific infrared analysers and mass spec-

45

trometer, Toutain et al. [20] monitored CO2 , He and 222 Rn in gases from a well located at the base of the active cone on Vulcano Island (Italy). Shimoike and Notsu [21] reported on a system to monitor volcanic gases extracted from an observation well located 3 km north of the summit of the Izu-Oshima Volcano (Japan) using IR sensors. In this paper we extend results and possibilities offered by an our preliminary approach to continuous hydrothermal/volcanic gas monitoring [22] and demonstrate the usefulness of a quadrupole mass spectrometer, deployed at the Pisciarelli fumarolic field (Campi Flegrei, Southern Italy) to improve present-day strategies of geochemical monitoring. Considering that data processing and interpretation are relevant part of the monitoring strategy, this paper shows how the described data acquisition and treatment brings advancements in our understanding of the investigated fumarolic system, included at one of the most dangerous volcanic areas on Earth. 2. Campi Flegrei setting 2.1. Generalities The restless Campi Flegrei caldera (CFc), often referred to as a ‘supervolcano’ [23] was formed by two major events: the Campanian Ignimbrite eruption, 39 ka ago [24,25], and the Neapolitan Yellow Tuff eruption, 15 ka ago [26]. More than 1.5 million people live within the CFc and surrounding areas, including the city of Naples. Volcanic risk has enormously increased through time as a consequence of rapid population expansion and strong, often irregular, urbanization. The previous eruption occurred in 1538 AD and was preceded by uplift episodes lasting decades [27]. The caldera resurgence is typically characterized by phases of uplift and subsidence over a range of timescales [28–31], not necessarily culminating in an eruption. The last important strong ground uplift (∼3.8 m) occurred from 1969 to 1984, followed by a period of relatively slow subsidence, interrupted by minor uplifts in 1989, 1994 and 2000 [32]. Since 2005 a renewed uplift phase affects the area, at a much smaller rate [33,34], resulting in a ground uplift of 38 cm in the last ten years [35]. Together with a moderate seismicity [34], such a ground deformation is accompanied by other signs of potential reawakening, particularly variations in the chemical composition of fumarolic emission, and flux gases degassing from soils [11,36–38]. These geochemical anomalies reveal a progressive increase in the CO2 -rich magmatic contribution to fumarolic fluids since 2000, largely prevailing over the meteoric/hydrothermal contribution [37,39,36]. The area is affected by intense, diffuse degassing and fumarolic activity in both Solfatara crater and Pisciarelli fault [5], placed close to the town of Pozzuoli. Detailed geochemical analyses of the fumaroles [37–39,36,40], coupled with the measurements of soil diffuse degassing [39–41] and with physical numerical simulations of the hydrothermal system [42], suggests that magma degassing episodes have a relevant role in triggering the volcanic unrest periods that periodically affect the area [37,43,44], as strongly supported by temporal coherence between changes in gas composition and uplift [11,45]. In January 2013, based on persistence of ground uplift and gas chemistry data, the Dipartimento della Protezione Civile (Italian Government) raised the state of the CFc from the green level (quiet) to the yellow level (scientific attention) [46]. 2.2. The fumarolic activity and state of art in geochemical monitoring at the Pisciarelli site The Pisciarelli area is a fault-related fumarolic field located a few hundred meters east of Solfatara crater (Fig. 1). This area, the most

46

A. Fedele et al. / International Journal of Mass Spectrometry 415 (2017) 44–54

Fig. 1. The Campi Flegrei caldera (CFc), Southern Italy. The unsteady Pisciarelli fumarolic field is located a few hundred meters east of the Solfatara crater. Both fumarolic fields display vigorous hydrothermal activity within a densely populated area.

Fig. 2. The Pisciarelli fumarolic field. In the foreground the main fumarolic vent and the bubbling mud pool. Our investigated, and ephemeral, emission points (in 2009 and 2012 campaigns) are shown.

active fumarolic vent aea in the entire Mediterranean volcanic district [8], is characterized by the presence of fractures and is affected by a consistent soil degassing and fluid emission (Fig. 2) from vents, mostly ephemeral and pools. Chiodini et al. [41] first identified a large diffuse degassing structure (DDS) with an anomalously high CO2 flux release, which also includes the Solfatara crater, a tuff cone near Pozzuoli that formed ∼3.9 ka BP [29]. The bulk CO2 flux from the Pisciarelli and Solfatara sites attains 50,000 g/m2 d, over an area of ∼1 km2 [44]. Fumarole discharges are mainly composed by H2 O (>70%) followed by CO2 and H2 S and have surface temperatures between 100 and 110 ◦ C [47]. A strong increase in fumarolic flow rate was reported to occur from 2005 to 2013 [40], also testified by continuous measurements of soil CO2 flux at the FLXOV3 station [35]. This degassing activity was episodically accompanied by local seismic activity [48] and by some macroscopic changes such as (1) the appearance of a vigorous degassing vent (March 2009), which is currently the main active gas source at this site (T > 110 ◦ C), (2) a mud bubbling pool (T ∼ 90 ◦ C). These changes were probably accompanied by a small

explosion [49] and the onset of a small active fumarolic field in the southern part of the area. In March 2009, CO2 flux was evaluated at only ∼18 tons/day; but, from 2012 to 2014 fumarolic emissions increased, ranging from ∼170 [5] to ∼380 tons/day [8]. During the recent ongoing unrest (in January 2013), a new vent appeared, emitting for few days high-pressure steam and liquid water up to 3–4 meters at least, partly linked to heavy rainfall during the last week of January. Aiuppa et al. [11] reported an increase of fumarolic CO2 output during the campaign carried out at the end of January 2013, with an unusually strong high (>3 m) jet of gas and hot water. In our experience at Pisciarelli, we never observed a constant pressure of gas discharged from investigated vents, but a pulsating behaviour. We observed many times, but never quantified, pressure increases at some vents and pressure drops at others, such that in some cases it was not possible to attain the minimal flow rates which are necessary for the “traditional” gas sampling by using the technique of Giggenbach and Goguel [50]. Consistent with these field observations, a nearly linear trend of the peak temperatures, oscillating from about 95 ◦ C up to around 115 ◦ C, has been recorded from 2003 to date [40]. Fig. 3 shows such temperature fluctuations at the main fumarolic outlet during the periods of QMS deployment. These testify to the unsteady nature of the fumarole, likely related to a still transient balance between the evolving and heating up reservoir conditions and the filtration of gravitative waters in the unsaturated subsoil of the fumarole. Such a behaviour, for example, is not observed at the two main nearby Solfatara fumaroles, which display stable temperatures (∼160 ◦ C at Bocca Grande and ∼145 ◦ C at Bocca Nuova; [40]), hence dynamically stable pathways of ascent fluids. We relate such a strong pulsating behaviour to the immaturity of the fumarolic ascent paths, and propose that strong fluctuations in the mass and energy balances of the fumarole are produced by the variable interaction of ascending fluids with the percolating surface waters, which are formed by a mix of rainwaters and condensed steam. In summary, the emission of gases and fluids at Pisciarelli is affected by near-surface secondary processes, probably of a seasonal nature that may mask the deep signals related to the temperature-pressure changes occurring in the hydrothermal system [49,51]. Most of the volcanic monitoring at the Pisciarelli fumarolic field has been so far carried out discontinuously. In particular, discontin-

A. Fedele et al. / International Journal of Mass Spectrometry 415 (2017) 44–54

47

Fig. 3. Temporal evolution of outlet temperatures. Chronogram of the Pisciarelli outlet temperatures during the 2009 and 2012 sampling campaigns. Also reported are temperature data taken from INGV-OV bulletins.

uous gas analysis at Pisciarelli started in 1999 [52] with a sampling frequency of less than one per month [35]. The flux of carbon dioxide dissipated in diffuse form via the soil has been characterized by Chiodini et al. [41]. In contrast, the gas output sustained by fumarolic activity was unconstrained (before the campaigns carried out from Aiuppa et al. [11] and Pedone et al. [5]), which reflects current technical inability to make gas flux estimates at low-temperature fumarolic fields. At these sites, where sulphur dioxide (SO2 ) is typically absent, conventional ultraviolet sensing techniques (e.g. [53]) are therefore useless, and a direct but challenging measurement of the CO2 emission rate is required [5,11]. The present-day status in the continuous monitoring of geochemical parameters at Pisciarelli is quite poor. The afore mentioned automatic station (FLXOV3) for the measurement of soil CO2 fluxes was installed for the first time in March 2002, but it was seriously and repeatedly damaged during the period 2002–2005; therefore, the data set for this period is largely incomplete [45]. The only attempt to determine automatically the composition of emitted gases from the fumaroles was carried by Aiuppa et al. [11] by means of a MultiGAS analyzer. Considering the visible, but not systematically quantified, increasing activity at the Pisciarelli site in the last decade, including small events of dark mud expulsion [11,45,49] and together with the highly pulsating behaviour in flow rates, it is now more important than ever to pursue continuous gas monitoring of such an unsteady and evolving fumarolic field in order to better understand precursors of the whole area unrest. Consequently, at the beginning of 2009, we started evaluating the feasibility of a continuous gas monitoring station, mainly equipped with a quadrupole mass spectrometer (QMS Pfeiffer OmnistarTM ), installed at the Pisciarelli site.

Moreover, the detection limit for H2 S (increasing to 200 ppmv) is due to the isobaric interferences with oxygen. The analyser of the qms is capable to record ion currents in the mass range 0–100 amu. Multiple gas concentrations detection was performed by analysing ion currents on peak maxima of eleven channels with the following mass over charge ratios m/z: 2, 4, 14, 15, 28, 32, 33, 36, 40, 44 and 84. The number of channels was reduced to eight in 2012 m/z: 2, 4, 14, 15, 32, 34, 40, 44. The detection time for each channel was set to 1 s, resulting in 11 respectively 8 s a complete record of all gases. Data were recorded minutely in order to reduce the data volume. The gas entrance into the quadrupole is ensured by its working pressure (lower than the atmospheric one); when the pressure is reduced, through a capillary (GSD 320 OmniStar, 2 m length), to ∼7 × 10−8 mbar, the gas can be pumped at ∼4 × 10−6 mbar value. In our study, the acceleration voltage was set to 90 eV in order to achieve an acceptable ion yield at a relatively low yield of multiply charged ions [22]. Multiple charged ions may cause significant isobaric interferences on mass-to-charge ratio m/z 20 (20 Ne and 40 Ar2+ ), m/z 22 (22 Ne and CO 2+ ). The resolution of the QMS does not 2 allow discriminating between 20 Ne and 40 Ar2+ respectively 22 Ne and CO2 2+ ; we therefore do not report neon concentrations here. The amount of hydrocarbons is generally low in volcanic gases with methane concentration in the water-free gas phase of Pisciarelli volcanic gases in the range of 10–100 ppmv. Isotopic gas equilibria studies suggest that CH4 is formed by reduction of CO2 [39] rather than by thermal degradation of organic matter. The amount of C2+ (ethane, propane, butane) in Pisciarelli volcanic gases is so low [54] that formation of fragmented and/or multiple charged ions is negligible. Methane was determined on m/z 15, with 85% yield, compared to m/z 16 to circumvent the problem of isobaric interferences with O2 2+ .

3. Gas mass spectrometry based on QMS: experimental field set-up

3.2. Calibration

3.1. Instrument During the surveys, carried out in 2009 and 2012, at the Pisciarelli site, a quadrupole analyser was used in dynamic mode, i.e. the gas was continuously pumped to the mass spectrometer, which was left to acquire (at 1 Hz), measuring the concentrations of gases (H2 , H2 S, He, Ar, CH4 , CO2 , N2 , O2 ) discharging from the vent. The analyzed species have different detection limits (e.g.